{"id":"W4411374363","doi":"10.1145/3722212.3725097","title":"Demonstrating CatDB: LLM-based Generation of Data-centric ML Pipelines","year":2025,"lang":"en","type":"article","venue":"","topic":"Advanced Data Storage Technologies","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Pipeline transport; Engineering; Mechanical engineering","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003214826,0.001688186,0.0006819353,0.001817328,0.0009141781,0.003249234,0.003679719,0.001343411,0.02119537],"category_scores_gemma":[0.01591095,0.001316103,0.001393647,0.00225529,0.001283924,0.004928643,0.005453703,0.002677456,0.01288352],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001726237,"about_ca_system_score_gemma":0.003170296,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01557747,"about_ca_topic_score_gemma":0.01513101,"domain_scores_codex":[0.9970238,0.0005060296,0.0003064116,0.0007685808,0.00118738,0.0002078941],"domain_scores_gemma":[0.9922044,0.00278776,0.0002195008,0.002964355,0.001509283,0.0003146724],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001478926,0.0004845244,0.006268806,0.001056588,0.000178378,0.0005281996,0.001244355,0.02446404,0.01865359,0.0197958,0.6647155,0.2611313],"study_design_scores_gemma":[0.000767156,0.0002996463,0.003357588,0.0002467267,0.00007253856,0.0004285277,0.0006081323,0.495822,0.1001607,0.03261332,0.3653636,0.0002600511],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"software","genre_gemma":"empirical","genre_scores_codex":[0.01757124,0.000508137,0.2656784,0.00132108,0.0005445201,0.0003517802,0.02466434,0.6753853,0.01397518],"genre_scores_gemma":[0.1431246,0.0005221281,0.6464314,0.002323392,0.0001339087,0.001192531,0.1206059,0.0730207,0.01264531],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02119537,"threshold_uncertainty_score":0.07090563,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07181776484659808,"score_gpt":0.3211208289744126,"score_spread":0.2493030641278146,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}